Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Yang, Xiguang, Arora, Krish, Bachmann, Michael
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2501.08341
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917892871684096
author Yang, Xiguang
Arora, Krish
Bachmann, Michael
author_facet Yang, Xiguang
Arora, Krish
Bachmann, Michael
contents We investigate the loss landscape and backpropagation dynamics of convergence for the simplest possible artificial neural network representing the logical exclusive-OR (XOR) gate. Cross-sections of the loss landscape in the nine-dimensional parameter space are found to exhibit distinct features, which help understand why backpropagation efficiently achieves convergence toward zero loss, whereas values of weights and biases keep drifting. Differences in shapes of cross-sections obtained by nonrandomized and randomized batches are discussed. In reference to statistical physics we introduce the microcanonical entropy as a unique quantity that allows to characterize the phase behavior of the network. Learning in neural networks can thus be thought of as an annealing process that experiences the analogue of phase transitions known from thermodynamic systems. It also reveals how the loss landscape simplifies as more hidden neurons are added to the network, eliminating entropic barriers caused by finite-size effects.
format Preprint
id arxiv_https___arxiv_org_abs_2501_08341
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Dissecting a Small Artificial Neural Network
Yang, Xiguang
Arora, Krish
Bachmann, Michael
Disordered Systems and Neural Networks
Statistical Mechanics
Machine Learning
Computational Physics
We investigate the loss landscape and backpropagation dynamics of convergence for the simplest possible artificial neural network representing the logical exclusive-OR (XOR) gate. Cross-sections of the loss landscape in the nine-dimensional parameter space are found to exhibit distinct features, which help understand why backpropagation efficiently achieves convergence toward zero loss, whereas values of weights and biases keep drifting. Differences in shapes of cross-sections obtained by nonrandomized and randomized batches are discussed. In reference to statistical physics we introduce the microcanonical entropy as a unique quantity that allows to characterize the phase behavior of the network. Learning in neural networks can thus be thought of as an annealing process that experiences the analogue of phase transitions known from thermodynamic systems. It also reveals how the loss landscape simplifies as more hidden neurons are added to the network, eliminating entropic barriers caused by finite-size effects.
title Dissecting a Small Artificial Neural Network
topic Disordered Systems and Neural Networks
Statistical Mechanics
Machine Learning
Computational Physics
url https://arxiv.org/abs/2501.08341